Arborescent Micelles: Dendritic Poly(γ-benzyl L-glutamate) Cores
Grafted with Hydrophilic Chain Segments
Greg Whitton and Mario Gauthier*
Supporting Information
EXPERIMENTAL PROCEDURES
Materials
Ethyl acetate (Caledon, 99+%) was stirred overnight with LiAlH4 (Aldrich, 95%) under N2 and
distilled immediately before use. Tetrahydrofuran (THF; Aldrich, ≥99%) for anionic
polymerization was distilled over sodium-benzophenone ketyl under N2. Toluene (Caledon, ACS
Reagent) for anionic polymerization was distilled over oligostyryllithium under N2. Ethylene oxide
(EO, Air Liquide) was purified with phenylmagnesium chloride as described below.
Quantification of Primary Amines by 19F NMR Analysis
The terminal primary amines on the linear polymers were quantified by a procedure adapted from
Ji et al.1 using
19
F NMR analysis. For example a linear PEO sample synthesized from 3-
aminopropanol, α-amino PEO110 (0.115 g, 2.2510-5 mol of chains), was dissolved in 3 mL of
deuterated DMSO (d6-DMSO). A solution of trifluorobenzaldehyde (TFBA, 0.1191 g, 6.8410-4
mol), and benzotrifluoride (BTF, 0.1014 g, 8.1510-4 mol) in 2 g of d6-DMSO was prepared (BTF
served as an internal standard). The reagent solution (0.2306 g, 7.0710-5 mol TFBA, 7.1810-5
mol BTF) was added to the polymer solution and stirred for 2 h. A 0.5 mL sample was transferred
1
to an NMR tube for analysis with 64 scans, and the integrated peak areas from the 19F NMR spectra
were used to determine the fNH2 values as described previously.2
Synthesis of 2,3-Epoxy-1-(1-ethoxyethoxy)propane (Glycidol Acetal)
The synthetic procedure used was as described by Fitton et al.3 2,3-Epoxypropanol (40.0 g, 0.54
mol) and ethyl vinyl ether (200 mL) were loaded in a 500 mL RBF with a magnetic stirring bar
and immersed in an ice-water bath. A catalytic amount of p-toluenesulfonic acid monohydrate (1.0
g, 5.3 mmol) was then added slowly, to ensure that the reaction temperature did not exceed 40 ºC
and to avoid the evaporation of ethyl vinyl ether. The reaction was removed from the ice bath and
allowed to warm to room temperature and to proceed for 3 h. A solution of saturated sodium
hydrogen carbonate was then added until the pH was slightly basic (approx. 100 mL). The organic
layer was isolated, dried over MgSO4, and concentrated under reduced pressure. Distillation of the
residue under reduced pressure gave the monomer as a colorless liquid that was stored under
nitrogen at 4º C. Yield: 61.5 g (78%); 1H NMR: (300 MHz, CDCl3) δ: 4.65 (q, 1H), 3.75-3.19 (m,
4H), 3.04 (m, 1H), 2.68 (m, 1H), 2.50 (m, 1H), 1.17 (d, 3H), 1.10 (t, 3H).
Synthesis of Diphenylmethylpotassium
The procedure used for the synthesis of diphenylmethylpotassium (DPMK) was adapted from
Normant and Angelo.4 A 3-neck RBF with a magnetic stirring bar was attached to a high-vacuum
line, flame-dried, and purged with nitrogen. Dry THF (150 mL) was added to the flask, followed
by potassium metal (4.26 g, 109.2 mmol, 2 eq) cut into small pieces and naphthalene (7.0 g, 54.6
mmol, 1 eq). The solution became dark green and was allowed to stir for 30 min. Diphenylmethane
(18.3 mL, 108.7 mmol, 2 eq) was then added to the flask with a syringe. The reaction was allowed
to proceed overnight to give a dark red DPMK solution that was stored at room temperature under
nitrogen.
2
Titration of the DPMK solution was performed using acetanilide under nitrogen. A 3-neck
RBF was attached to the high vacuum line, flame-dried, and purged with nitrogen. THF (30 mL)
was added, followed by a few drops of DPMK solution, until the solution remained a pale yellow
color. Acetanilide (53.0 mg, 0.39 mmol) was added to the RBF, at which point the color
disappeared. The DPMK solution was slowly added (0.77 mL) to obtain the same pale yellow
color present initially. This volume corresponded to a DPMK concentration of 0.51 M.
Synthesis of α-Amino Poly(glycidol acetal) (Amino-PGlyAc62)
In typical anionic polymerization procedures, the monomer is purified on a high-vacuum line and
transferred to an ampoule immediately before use. The glycidol acetal monomer cannot be purified
by that technique due to its high boiling point of 152-154 ºC however.3 It was rather distilled over
triisobutylaluminum in a fractional vacuum distillation setup directly before use. Glycidol acetal
(40.0 g) was placed in a 100 mL RBF with a stirring bar and purged with nitrogen.
Triisobutylaluminum (2 mL, 2 mmol) was added to the flask with stirring. The flask became warm
within minutes of adding the triisobutylaluminum. After the flask had cooled to room temperature
the glycidol acetal was distilled under reduced pressure into a RBF that was then sealed with a
rubber septum and purged with nitrogen.
A 3-neck RBF with a stirring bar was attached to the vacuum line, flame-dried under high
vacuum, and purged with nitrogen. Dry THF (25 mL) was added to the RBF, followed by DPMK
drop-wise until a faint yellow color persisted in the solution. 3-Aminopropanol (0.19 mL, 2.53
mmol) was then added, followed by DPMK (5.1 mL, 0.51 M) to deprotonate the alcohol. The
solution became milky, but DPMK was added further until the solution maintained a faint
yellow/red color for one minute. Freshly distilled glycidol acetal (25.2 g, 0.173 mmol, target Xn =
68, Mn = 10,000 g/mol) was added and the flask was sealed. The temperature was increased to 65
3
ºC using an oil bath and the reaction was left stirring overnight under nitrogen. Degassed acidified
methanol was then added to terminate the reaction. The solution was transferred to a regenerated
cellulose dialysis bag with a 1000 molecular weight cutoff (MWCO) and left to stir in THF. The
THF bath was changed once after 3 h and left to stir overnight. The dialysis bag was then emptied
into a RBF and the THF was evaporated under vacuum to give a reddish-brown viscous polymer.
Yield: 16.4 g (65%). SEC (THF): Mn = 9100 g/mol (Xn = 62), Mw/Mn = 1.08; 1H NMR: (300 MHz,
CDCl3) δ: 4.66 (q, 1H), 3.70-3.39 (m, 7H), 1.24 (d, 3H), 1.15 (t, 3H) (initiator fragment protons
not detectable).
Synthesis of α-Azido PGlyAc220
To obtain a high molar mass poly(glycidol acetal) sample (target Mn = 30,000 g/mol), a different
polymerization method was necessary to avoid a high dispersity due to chain transfer reactions.5,6
The procedure was adapted from Gervais et al.6 The initiator tetrabutylammonium azide (0.38 g,
1.33 mmol) was dried before use by three cycles of azeotropic distillation with dry toluene under
vacuum, and stored under nitrogen after redissolution in 20 mL of toluene in a glass ampoule
sealed with a Teflon stopcock. A 1-L, 5-neck RBF was evacuated under high-vacuum, flame-dried,
and purged with nitrogen. Dry toluene (400 mL) was then added and the RBF was cooled to -30
ºC with dry ice in a 2-propanol/water bath. Glycidol acetal (40.0 g, 0.274 mol, target Xn = 206, Mn
= 30,000 g/mol, freshly distilled over triisobutylaluminum), the initiator solution, and
triisobutylaluminum (5.9 mL of solution, 5.9 mmol) were added in succession, and the 2propanol/water bath was removed to allow the temperature to increase to room temperature. The
reaction was allowed to proceed overnight, and degassed acidified methanol was added to
terminate the reaction. The toluene solution was concentrated to approximately 50 mL, transferred
to a regenerated cellulose dialysis bag with a 1000 MWCO, and left to stir in THF. The THF bath
4
was changed once after 3 h and left to stir overnight. The dialysis bag was then emptied into a RBF
and the THF was evaporated to give a clear viscous polymer. Infrared spectroscopy served for the
qualitative analysis of terminal azide groups, on a Bruker Vector 22 FT-IR spectrometer with the
OPUS 6.0 software to acquire and manipulate the spectra. The analysis was performed with 64
scans from 400 to 4000 cm-1 at 1 cm-1 resolution, with the viscous polymer placed directly between
NaCl plates. Yield: 34.5 g (86%). IR: -N3 stretch at 2102 cm-1. SEC (THF): Mn = 32,400 g/mol
(Xn = 220), Mw/Mn = 1.19; 1H NMR (300 MHz, CDCl3) δ: 4.66 (q, 1H), 3.67-3.35 (m, 7H), 1.24
(d, 3H), 1.15 (t, 3H).
Reduction of α-Azido PGlyAc220 to α-Amino PGlyAc220
Reduction of the α-azide to an α-amine group was done by loading α-azido PGlyAc220 (30.0 g)
into a 1-L RBF with 300 mL of THF under nitrogen, and adding a solution of lithium aluminum
hydride (LiAlH4, 6.0 g in 200 mL of THF). The reaction was refluxed for 3 h, and left at room
temperature overnight. Water (10 mL) was finally added slowly to destroy excess LiAlH4. The
solution was centrifuged to remove salts and placed in a 1000 MWCO regenerated cellulose
dialysis bag in THF overnight, producing a clear reddish brown viscous polymer. Yield: 24.0 g
(80%). IR: disappearance of -N3 stretch at 2102 cm-1. 1H NMR: (300 MHz, CDCl3) δ: 4.66 (q,
1H), 3.67-3.35 (m, 7H), 1.24 (d, 3H), 1.15 (t, 3H).
Ethylene Oxide Purification
Caution: EO is highly toxic and volatile (b.p. 10 ºC), so it should be manipulated with great care
in a well-ventilated fume hood, and the pure monomer should be cooled as much as possible to
avoid excessive pressure buildup. Ethylene oxide (EO) was purified on a high-vacuum line using
a manifold with connections for the EO tank line, and an ampoule containing a Teflon stopcock, a
magnetic stirring bar, and approximately 2 g of calcium hydride as a drying agent. The manifold
5
and the ampoule were evacuated and flame-dried, and EO (approximately 100 g) was condensed
under vacuum to the ampoule by cooling it in liquid nitrogen. The ampoule was then mounted on
another vacuum manifold equipped with an RBF containing a magnetic stirring bar, and another
ampoule with a Teflon stopcock. The EO was degassed with three successive freeze-pump-thaw
cycles. After closing the ampoule containing the EO, the rest of the manifold was evacuated and
flame-dried. After purging the apparatus with nitrogen, phenylmagnesium chloride solution
(PhMgCl, 9 mL, 2.0 M in THF) was added to the RBF on the manifold with a syringe. The THF
was removed under vacuum and ca. 15 g of EO was transferred to the RBF containing the PhMgCl.
The monomer was stirred for 1 h in an ice bath before slowly transferring it over to the empty
storage ampoule. The amount of EO transferred was 12.3 g.
Polymerization of EO
A 5-neck 500-mL RBF with a magnetic stirring bar was attached to the high-vacuum line with the
sealed ampoule containing the purified EO monomer (cooled with dry ice). The RBF was
evacuated, flame-dried, purged with nitrogen, and dry THF (120 mL) was added followed by
DPMK solution drop-wise until a faint yellow color persisted in the solution. 3-Aminopropanol
(0.19 mL, 2.53 mmol) was then added, followed by DPMK (5.1 mL, 0.51 M) to deprotonate the
alcohol. The solution became milky, and DPMK was added further until a faint yellowish-red color
persisted for 1 min in the solution. The EO monomer (12.3 g, 0.279 mol, target Xn = 110, Mn =
5000 g/mol) was then transferred under vacuum, the ampoule was sealed and the temperature was
brought to 45 ºC with an oil bath. The reaction was allowed to proceed for 6 d, after which time a
dark brown solution was obtained. Degassed acidified methanol was added to terminate the
reaction. The solution was concentrated to approximately 50 mL under vacuum and precipitated
in cold diethyl ether. A brown powder was recovered by suction filtration. It was redissolved in
6
methanol and precipitated in cold diethyl ether, recovered by suction filtration, and dried under
vacuum overnight to produce an off-white powder. Yield: 8.5 g (69%). SEC (DMF): Mnapp = 6200
g/mol, Mw/Mnapp = 1.16. 1H NMR: (300 MHz, CDCl3): Xn = 114, δ: 3.87-3.37 (m, 456H), 2.88 (br,
1H), 1.96 (br, -OH).
Synthesis of -tert-Butyl L-Glutamic Acid N-Carboxyanhydride (tBuGlu-NCA)
The procedure used was similar to the one reported previously for the synthesis of -benzyl Lglutamic acid N-carboxyanhydride.2 γ-tert-Butyl L-glutamic acid (10.0 g; 49.2 mmol) was
suspended in 300 mL of dry ethyl acetate in a 1-L RBF fitted with a refluxing condenser and a gas
bubbler. The flask was purged with N2 and heated to reflux. Triphosgene (5.6 g, 18.7 mmol) was
then added and refluxing was continued for 3 h. The flask was removed, stoppered, and cooled in
a freezer (–10 C) for 1 h. The solution was transferred to a cold separatory funnel and quickly
washed successively with 100 mL of ice-cold water and 100 mL of chilled 0.5% aqueous NaHCO3
solution. The organic phase was dried over anhydrous MgSO4, filtered, and concentrated to ca.
100 mL on a rotary evaporator. An equal volume of cold (–10 C) hexane was then added to induce
crystallization of the product. The mixture was left in the freezer overnight and the solid product
was recovered by filtration in a Schlenk funnel under N2. It was then dried overnight under vacuum
to yield a white powder, and stored under N2 in a freezer (-20 C).Yield = 8.6 g (76 %). 1H NMR
(300 MHz, CDCl3) δ: 6.75 (s, 1H), 5.11 (s, 2H), 4.38–4.33 (t, 1H), 2.59–2.53 (t, 2H), 2.35–2.21
(m, 1H), 2.21–2.02 (m, 1H), 1.42-1.37 (s, 9H).
Polymerization of tBuGlu-NCA
The procedure used was similar to the one reported previously for the polymerization of -benzyl
L-glutamic
acid N-carboxyanhydride.2 The tBuGlu-NCA monomer (1.87 g, 8.1 mmol) was
7
dissolved in dry DMF (15 mL) in a 100-mL RBF at 0 C and n-hexylamine (50 L, 0.38 mmol,
for a target Xn = 20) was added with rapid stirring. The reaction was allowed to proceed for 5 d at
0 C. The linear polymer was recovered by precipitation in diethyl ether, suction filtration, and
drying under vacuum overnight to give a white powder. Yield = 0.8 g (53%). SEC (DMF): Mw/Mn
= 1.15. 1H NMR (300 MHz, d6-DMSO): Xn = 11.8, δ: 4.19 (br, 12H), 2.24-2.18 (br, 24H), 1.811.69 (m, 24H), 1.34 (s, 108H), 1.25-1.13 (b, 10H), 0.81–0.79 (t, 3H).
RESULTS AND DISCUSSION
Synthesis of α-amino-PGlyAc62
The procedure for the synthesis of α-amino-PGlyAc62 was adapted from Dworak et al.7 using the
initiator 3-aminopropanol to produce poly(glycidol acetal) with a terminal primary amine
functionality. 3-Aminopropanol was deprotonated with DPMK to produce an alcoholate capable
of initiating the polymerization of glycidol acetal. The amine protons of the initiator molecule do
not disrupt the polymerization reaction, as they are orders of magnitude less labile than the proton
from the alcohol group on the initiator. For comparison 1-propanol has a pKa of 16, whereas
primary amines have a pKa of ca. 36. Therefore the abstraction of a proton from the amine group
by the alcoholate functionality is negligible.
The anionic polymerization of unprotected glycidol leads to a branched polymer structure due
to the fast exchange between the alcoholates and the hydroxyl groups present in the polymer chain.
This causes significant branching, which has occasionally proven to be useful.8 To obtain a linear
polymer with a narrow molar mass distribution, a protected form of glycidol must be used. The
procedure developed by Fitton et al. 3 yields a protected glycidol monomer in the acetal form,
8
suitable for anionic polymerization, in high yield and purity. The polymerization of glycidol acetal
with the 3-aminopropanol/DPMK initiator system, to obtain linear PGlyAc with a primary amine
terminal functionality, is depicted in Scheme S1.
O
O
THF
H 2N
O
H 2N
OH
- +
H 2N
C 3H 6
O
H
OK
65o C, 16 h
DPMK
O
n
O
Scheme S1. Polymerization of glycidol acetal with 3-aminopropanol and DPMK.
Due to the high boiling point of the glycidol acetal monomer (152-154 ºC),3 it could not be
dried using high-vacuum line purification techniques often employed for monomers in anionic
polymerization. To achieve the high level of monomer purity required, purification was first
attempted by simple distillation without additives immediately before the reaction in a reduced
pressure distillation setup. Polymerization reactions using that monomer gave molar mass
distributions broader than typically expected for anionic procedures (Mw/Mn ≥1.20). An additional
attempt to distill the monomer from calcium hydride prior to use also had little influence on the
dispersity. Triisobutylaluminum was finally explored as a drying agent, as it is known to act as
proton scavenger9 and is safe to use with glycidol acetal since it has been employed in the activated
monomer polymerization of glycidol acetal.6 This purification technique yielded much better
results with respect to the dispersity, with Mw/Mn ≤ 1.10.
Synthesis of α-Amino PGlyAc220
The synthesis of α-azido PGlyAc220 relied upon an activated oxirane monomer polymerization
method introduced by Billouard et al.5 using azide salt initiators along with glycidol acetal. The
9
activated monomer polymerization mechanism initiated by an azide salt is depicted in Scheme S2.
This polymerization technique allows the synthesis of poly(glycidol acetal) with Mn up to 30,000
g/mol, while maintaining a relatively low dispersity (Mw/Mn ≤ 1.30).
The azide functionality of α-azido poly(glycidol acetal) was reduced to a primary amine using
LiAlH4 in THF. The reduction was easily monitored by infrared (IR) analysis, since azide
stretching vibrations produce a strong absorption near 2100 cm-1. The disappearance of this peak
should therefore be indicative of the presence of a primary amine at the chain end, even if the
molar mass of the polymer is relatively high. This is seen in Figure S1, by comparing IR spectra
obtained before and after the reduction reaction. To ensure that no degradation occurred during the
reduction, SEC measurements were compared before and after the reaction yielding Mw/Mn values
of 1.19 and 1.20, respectively.
O
N3
- +
O
NBu 4
+
Al(i-Bu) 3
(4.5 eq.)
[
N3 ---Al(i-Bu) 3
O
]Toluene
+

+
Al(i-Bu) 3
O
25o C, 3 h
N3
O
H
n
O
O
O
O
Scheme S2. Activated monomer mechanism for the anionic polymerization of glycidol acetal
initiated by tetrabutylammonium azide.
10
Figure S1. IR Spectra for α-azido PGlyAc220 (top) and α-amino PGlyAc220 (bottom).
1
H NMR analysis served to confirm that no deprotection occurred during the synthesis and
the isolation of the PGlyAc samples. The initiator protons from 3-aminopropanol (for PGlyAc62)
and the protons  to the azide or amine functionality (for PGlyAc220) were not resolved from the
repeating units however, so no Mn values could be derived from 1H NMR analysis.
Synthesis of α-Amino PEO110
Poly(ethylene oxide) is well-known for being biocompatible and water-soluble, so it is a natural
choice as a hydrophilic shell material for the arborescent copolymers. Different approaches are
available to obtain primary amine-terminated poly(ethylene oxide). Commercially available PEO
with a terminal hydroxyl group can be modified through either halogenation10 or tosylation11 of
the hydroxyl group, which can then be converted to an azide functionality. Reduction of the azide
yields a terminal primary amine functionality. Since this approach involves multiple reaction steps,
the probability of incomplete or side reactions leading to lower levels of amine functionality is
increased. It therefore seemed more practical to make use of the anionic polymerization of ethylene
oxide with a bifunctional initiator to ensure a high level of amine functionality. 1H NMR spectra
11
for a commercial PEO monomethyl ether sample with Mn = 5000 g/mol and the α-amino PEO110
sample synthesized are compared in Figure S2. For α-amino PEO110, a number-average degree of
polymerization (Xn) of 110 was calculated from the integrated intensities for the -CH2- protons
next to the terminal amine ( 2.9 ppm) relatively to the four protons in the repeat units. This
corresponds to a Mn = 5100 g/mol. The peak near  2.8 ppm is due to residual diethyl ether.
Figure S2. 1H NMR Spectra for PEO monomethyl ether with Mn = 5000 g/mol (top) and
synthesized α-amino PEO110 (bottom) in CDCl3.
Determination of Primary Amine Functionality (fNH2) of Linear Polymers
The primary amine functionality level, fNH2, was also reported in Table 1 of the paper. An example
for the determination of fNH2 is provided for α-amino PEO110. Following a procedure developed by
Ji et al.1 the terminal amine of -amino PEO110 was reacted with trifluorobenzaldehyde (TFBA)
to determine the fraction of polymer chains containing a primary amine functionality. The same
12
procedure was applied previously to the analysis of linear PBG.2 The amine quantification reaction
with TFBA is described in Scheme S3. The reaction produces an imine that changes the
environment of the fluorine atoms on the benzene ring of TFBA. A 19F NMR spectrum obtained
for the PEO110 amine quantification reaction using TFBA and BTF is shown in Figure S3. Three
molar equivalents of TFBA were used to ensure that the PEO110 chains reacted completely with
TFBA. Three equivalents of BTF were also used as internal standard in the measurement.
H
O
O
n
O
(C3H6) NH2
CF3
CF3
BTF
+
d6-DMSO, 2 h
H
O
CF3
O
n
(C3H6)
N
H
Scheme S3. Reaction of -amino PEO110 with trifluorobenzaldehyde to produce an imine
quantified by 19F NMR spectroscopy.
Figure S3. 19F NMR spectrum for amine quantification in -amino PEO110.
Equation S1 shows how fNH2 was determined by integration of the peaks in the
19
F NMR
spectrum. Since the number-average molar mass of the polymer is known (Mn = 5100 g/mol), the
number of moles of chains in the sample is known. The integration value for the imine fluorides
13
(0.287) was thus compared to the integration value for the BTF fluorides (internal standard, 1.000)
to give fNH2 = 0.91. Since there are errors involved in determining the exact Mn value for PEO110
by 1H NMR analysis and in the amine quantification procedure, the fNH2 value also has an
uncertainty associated with it. To take this into account, an excess of side chains was used in the
coupling reactions. The
19
F analysis rather served to confirm that terminal primary amine
functionalities were indeed present on most polymer chains, and therefore that the anionic
polymerization reaction using 3-aminopropanol as bifunctional initiator was successful. The same
amine quantification procedure was carried out for PGlyAc62 and PGlyAc220. Sample PGlyAc62
produced an fNH2 = 0.64; however, PGlyAc220 did not produce a detectable signal for the imine
even for a reaction time of 4 h, likely due to the larger chain length of the polymer. The PGlyAc220
polymer was still successfully used in the grafting reactions even though no primary amine
quantification was possible for that sample.
𝒎𝒐𝒍 𝒊𝒎𝒊𝒏𝒆
𝒇𝑵𝑯𝟐 = 𝒎𝒐𝒍 𝒑𝒐𝒍𝒚𝒎𝒆𝒓 𝒄𝒉𝒂𝒊𝒏𝒔 =
𝑰𝒎𝒊𝒏𝒆 𝒊𝒏𝒕𝒆𝒈𝒓𝒂𝒕𝒊𝒐𝒏
)
𝑩𝑻𝑭 𝒊𝒏𝒕𝒆𝒈𝒓𝒂𝒕𝒊𝒐𝒏
𝒈 𝒐𝒇 𝒑𝒐𝒍𝒚𝒎𝒆𝒓
(
)
𝑴𝒏 𝒐𝒇 𝒑𝒐𝒍𝒚𝒎𝒆𝒓
𝒎𝒐𝒍 𝑩𝑻𝑭×(
=
𝟎.𝟐𝟖𝟕
)
𝟏.𝟎𝟎
𝟔.𝟒𝟎×𝟏𝟎−𝟓 𝒎𝒐𝒍×(
𝟎.𝟏𝟎𝟑𝟎 𝒈
𝟓𝟏𝟎𝟎 𝒈/𝒎𝒐𝒍
= 𝟎. 𝟗𝟏
(S1)
Synthesis of Poly(γ-tert-butyl L-glutamate)
Cleavage of the tert-butyl ester protecting group from that polymer yields poly(L-glutamic acid)
(PGA), known to be a biocompatible water-soluble material at neutral or basic pH, since the
carboxylic acid units have a pKa close to 5. PGA has indeed been used in biomedical applications12
and should also be useful as hydrophilic shell material for the PBG micelles.
There is very little published work on the synthesis of tBuGlu homopolymers. Ngyuen et al.
reported the synthesis of PtBuGlu from the NCA monomer in chloroform at 0º C for up to one
14
week.13 Dispersities of 1.34 and 1.16 were reported for polymers with similar Xn = 63 and 59,
respectively, which points at irreproducibility issues.
The synthesis of PtBuGlu was attempted by a procedure similar to the one described for linear
PBG2 but unfortunately, the polymerization did not proceed as expected under these conditions.
The ring-opening polymerization of tBuGlu-NCA, shown in Scheme S4, used n-hexylamine with
the monomer in DMF at 0º C under nitrogen for 5 days. The polymer yield was low due to
incomplete monomer conversion, even though the reaction was allowed to proceed for 2 days
longer than in the corresponding linear PBG syntheses. While the target Xn was 20, Xn = 12 was
obtained experimentally. The dispersity of the sample nevertheless remained low (Mw/Mn = 1.15),
indicating that the polymerization was not affected by side reactions. It was also determined that
fNH2 = 0.99 for PtBuGlu12. To produce a water-soluble polymer, the tert-butyl ester was removed
by dissolving it in pure TFA for a few minutes. This reaction also served for the selective
deprotection of the shell material after linear PtBuGlu was grafted onto the PBG substrates, to
produce poly(L-glutamic acid) (PGA) segments soluble in physiological buffer (pH 7.4).
O
O
n-C6H13 NH2
+
O
N
H
O
DMF
0 oC, 5 days
O
H
N
n-C6H13 N
H
NH2
O
n
O
O
O
O
O
Scheme S4. ROP of tBuGlu-NCA using n-hexylamine as initiator.
Determination of the Number-average Degree of Polymerization of (OtBu)2-PBG
The integrated peak intensity for the 18 protons of the tert-butyl ester groups in the initiator
fragment, found at 1.3 ppm, were compared with that for the two benzylic protons on each repeat
15
unit at 5.0 ppm. Residual DMF from the polymerization reaction is responsible for the peaks at
2.7, 2.9, and 7.9 ppm, and water from the d6-DMSO also appears at 3.3 ppm.
Figure S4. 1H NMR spectrum for Glu(OtBu)2·HCl-initiated linear PBG in d6-DMSO.
Confirmation for Removal of the tert-Butyl Ester Protecting Groups from PBG
More or less complete disappearance of the tert-butyl protons at 1.3 ppm is observed, and the small
peaks remaining near 1.3 ppm are from the n-hexylamine residues attached to the PBG substrates
in the previous grafting reactions. Comparison of the methine proton signal at 3.8-4.4 ppm to the
benzyl ester protecting group protons at 5 ppm before and after acidolysis was performed to ensure
that there was no significant loss of the benzyl ester protecting groups during acidolysis.
Figure S5. 1H NMR spectra in d6-DMSO for a G1 chain-end functionalized PBG sample before
(top) and after acidolysis (bottom) of the tert-butyl ester protecting groups.
16
Figure S6. SEC traces in DMF with 0.1% LiCl for (from top to bottom) G2PBG-g-PGlyAc220,
G1PBG-g-PGlyAc220, G2PBG-g-PGlyAc62, and G1PBG-g-PGlyAc62 arborescent copolymers.
Acidolysis of Acetal Protecting Groups on Arborescent PGlyAc Copolymers
The acidolysis of the acetal protecting group is represented in Scheme S5. SEC traces obtained for
the crude products obtained by different deprotection protocols are compared in Figure S5 for
sample G1PBG-g-PGlyAc220. The SEC trace for the purified copolymer before deprotection is also
provided in Figure S5 for comparison, to help determine the extent of degradation occurring for
the arborescent systems in each case.
PBG
NH
C3 H6
O
H
n
H
+
O
PBG
NH
C3H6
O
H
n
OH
O
Scheme S5. Acidolysis of the acetal groups in arborescent PBG-g-PGlyAc copolymers.
17
Figure S7. SEC traces in DMF with 0.1% LiCl for different deprotection reactions of G1PBG-gPGlyAc220. From top to bottom, G1PBG-g-PGlyAc220 (reference trace), AlCl3 deprotection,
formic acid deprotection, HCl deprotection (30 min), HCl deprotection (90 min).
The first deprotection method, reported by Namboodri and Varma,14 used a catalytic amount
of aluminum chloride hexahydrate in methanol to remove the acetal protecting groups from small
molecules, and was more recently applied by Dimitrov et al.15 to copolymers containing glycidol
acetal. The reaction, when allowed to proceed for 30 min, led to significant degradation according
to the SEC trace in Figure S5. The second deprotection method, with neat formic acid, was
proposed by Taton et al.16 for the deprotection of a poly(glycidol acetal)-b-poly(ethylene oxide)
copolymer. The first step involves dissolution of the polymer in neat formic acid to produce
formate groups, which are then saponified with potassium hydroxide in a dioxane/methanol
mixture to release the hydroxyl functionalities. This technique, when applied to the arborescent
copolymer, also led to significant degradation, as seen for the third SEC trace in Figure S5. More
recently, Mendrek et al. also investigated different techniques for the deprotection of poly(glycidol
acetal).17 They found that a concentrated HCl solution in DMF achieved 95% acetal group
cleavage after only 30 min, and 100% removal after 45 min. The fourth and fifth SEC traces in
Figure S5 are for the arborescent copolymers obtained under these conditions (also reported in the
experimental section), where the arborescent copolymer was dissolved in a DMF solution
18
containing 4.5 eq of HCl relatively to the acetal protecting groups. Complete removal of the acetal
protecting groups was confirmed by 1H NMR analysis after 30 min, although a small amount of
side chain degradation was also observed as a broad peak at elution volumes around 11.5-12 mL.
To confirm that this degradation did not affect the PBG substrate, the same experiment was run
with the G1PBG substrate in the HCl/DMF solution. SEC samples removed after up to 120 min
displayed no sign of degradation. It is therefore clear that a 30 min reaction time using HCl/DMF
works best for the deprotection of the arborescent copolymers. The small amount of degradation
products observed is likely due to random cleavage of some polyglycidol (PGly) segments. These
chain segments should not have a significant impact on the solution properties of the arborescent
copolymers, but they were nevertheless removed by dialysis of the crude product against methanol
in a 12,000-14,000 MWCO bag. The SEC elution curves obtained for the purified arborescent
PBG-g-PGly copolymers, shown in Figure S7, follow a similar trend to the elution curves for the
protected polymers displayed in Figure S6. SEC traces for the other copolymers synthesized with
PGlyAc, PGly, PEO and PtBuGlu side chains are provided in Figures S8-S12.
Figure S8. SEC traces in DMF with 0.1% LiCl for (from top to bottom) G2PBG-g-PGly220,
G1PBG-g-PGly220, G2PBG-g-PGly62, and G1PBG-g-PGly62 arborescent copolymers.
19
Figure S9. SEC traces in DMF with 0.1% LiCl for (from top to bottom) G3PBG-g-PEO110,
G2PBG-g-PEO110, and G1PBG-g-PEO110 arborescent copolymers.
Figure S10 SEC traces in DMF with 0.1% LiCl for (from top to bottom) G3PBG-g-PtBuGlu12,
G2PBG-g-PtBuGlu12, and G1PBG-g-PtBuGlu12 arborescent copolymers.
Figure S11 SEC traces in DMF with 0.1% LiCl for purified arborescent copolymers: (top to
bottom) G3PBG-eg-PGlyAc62, G2PBG-eg-PGlyAc62, and G1PBG-eg-PGlyAc62.
20
Figure S12 SEC traces in DMF with 0.1% LiCl for the purified arborescent copolymers: (top to
bottom) G3PBG-eg-PEO110, G2PBG-eg-PEO110, and G1PBG-eg-PEO110.
Dimension Calculations for the Chains Added in the Shell
The hydrophilic polymer chains forming the shell of arborescent copolymer micelles could have a
coiled, partly extended, or even fully extended chain conformation. However it should be
considered that, given the large entropic penalty resulting from a conformation change from a
coiled to an extended chain, the shell chains are unlikely to be significantly extended. It was indeed
demonstrated in a previous investigation on arborescent polystyrene end-grafted with a shell of
poly(ethylene oxide) chains that the increases in dh observed upon addition of the shell material
were consistent with the PEO chains having a coiled conformation.18 It will be shown below that
for the copolymers synthesized in the current investigation, if the chains are assumed to have a
coiled conformation, the increase in dh expected is on the order of 10 nm or less if the molecules
are not aggregated.
The coiled dimensions of polymers having the same molar mass as the side chains grafted
in the shell of the copolymers can be estimated from the intrinsic viscosity [], ideally obtained
under the same solvency conditions (solvent, temperature) used in the dynamic light scattering
(DLS) measurements, using a rearranged form of the Einstein equation
21
 24 M 

d h  
 10 N A 
1/ 3
(S1)
where M is the molar mass and NA is Avogadro’s constant. The intrinsic viscosity may be either
measured directly or estimated from the Mark-Houwink-Sakurada (MHS) equation [] = kM ,
where k and  are the experimentally determined MHS parameters.
Another method to estimate the dimensions of polymer chains in through the determination
of the diffusion coefficient Do of the molecules in DLS measurements, from which the
hydrodynamic diameter can be calculated from the Stokes-Einstein equation as
dh 
2kT
6Do
(S2)
where k is Boltzmann’s constant, T is the absolute temperature, and  is the viscosity of the solvent.
The variation in dh with molar mass may then be described through a scaling relation dh = kM
a
where the parameters k and a are determined experimentally.
In practice, the amount of information available on the intrinsic viscosity or the
hydrodynamic diameter of polymers under the same conditions used for the DLS measurements
tends to be rather limited. The values thus obtained should therefore be viewed as very
approximate, but are deemed sufficient for the purpose of the present discussion.
Hydrodynamic diameter data for linear polyglycidol in water were thus only reported19 for
a range of molar masses from Mw = 1 × 105 to 2.4 × 106 g/mol, while the PGly62 side chains used
in the present work had Mn = 5100 g/mol. Assuming that the scaling relation dh (nm) =
0.0278 Mw0.542 is still valid when extrapolated to M = 5100 g/mol, dh  h = 2.8 nm is estimated
for the shell thickness. A similar calculation for the PGly220 side chains yields h = 3.6 nm. Due to
22
the low molar mass of the M = 5100 g/mol side chains extrapolation of the scaling relation provided
in Reference 19 may be inaccurate, since short chains are expected to a have more rod-like
conformation in solution. To confirm this, DLS measurements were performed for the
(deprotected) PGly chain segments with M = 5100 g/mol to determine their hydrodynamic
diameter directly and yielded dh = 3.9 nm, indeed larger than the extrapolated value (dh = 2.8 nm).
Therefore a total increase in diameter of 2h = 7.8 nm is expected if the PGly62 chains are coiled.
Given the larger than expected size of the PGly62 side chains, a similar trend would be expected
for the PGly220 side chains. Consequently, a size increase somewhat larger than 2h = 2 × 3.6 nm
= 7.2 nm, i.e. around 10 nm would be expected when adding a shell of PGly220 chains.
The intrinsic viscosity of PEO110 was reported to be [] = 0.152 dL/g in water at 25oC.20
This parameter can be combined with M = 5000 g/mol in equation S1 above to obtain dh = 2.1 nm
(or 2h = 4.2 nm) for PEO110 side chains in the shell of the copolymers.
No information was found on the dimensions of short PGA chains in water, but if the
dimensions of one amino acid residue are taken as about 3.5 Å (in analogy to the fully extended
chains found in -sheets) with Xn = 12 for the PGA12 chains, rmax = 4.2 nm or 2 rmax = 8.4 nm is
obtained, which would represent the maximum increase in diameter possible upon addition of a
shell of PGA12 chains.
23
CONTIN Analysis Results for the End-grafted Copolymer Samples
100
100
G2PBG-eg-PGly62
60
40
G3PBG-eg-PGly62
80
Intensity (%)
Intensity (%)
80
20
60
40
20
0
0
10
100
10
100
Diameter (nm)
Diameter (nm)
100
100
G2PBG-eg-PEO110
60
40
G3PBG-eg-PEO110
80
Intensity (%)
Intensity (%)
80
20
60
40
20
0
0
10
100
10
100
Diameter (nm)
Diameter (nm)
100
G2PBG-eg-PGA12
Intensity (%)
80
60
40
20
0
10
100
Diameter (nm)
Figure S13. CONTIN analysis results for G2PBG and G3PBG end-grafted samples dispersed in
PBS solutions.
24
REFERENCES
1. S. Ji, T. R. Hoye, C. W. Macosko, Macromolecules 2005, 38, 4679-4686.
2. G. Whitton, M. Gauthier, J. Polym. Sci., Part A: Polym. Chem. 2013, 51, 5270-5279.
3. A. O. Fitton, J. Hill, D. E. Jane, R. Millar, Synthesis 1987, 12, 1140-1142.
4. H. Normant, B. Angelo, Bull. Soc. Chim. France 1960, 354-359.
5. C. Billouard, S. Carlotti, P. Despois, A. Deffieux, Macromolecules 2004, 37, 4038-4043.
6. M. Gervais, M. Labbé, S. Carlotti, A. Deffieux, Macromolecules 2009, 42, 2395-2400.
7. A. Dworak, I. Panchev, B. Trzebicka, W. Walach, Macromol. Symp. 2000, 153, 233-242.
8. A. Sunder, R. Hanselmann, H. Frey, R. Mülhaupt, Macromolecules 1999, 32, 4240-4246.
9. S. Wang, US Pat. 20050070675, 2005.
10. H. Gehrhardt, M. Mutter, Polym. Bull. 1987, 18, 487-493.
11. C. Yaniç, M. W. Bredenkamp, E. P. Jacobs, S. C. Spies, P. Swart, J. Appl. Polym. Sci. 2000,
78, 109-117.
12. C. Li, Adv. Drug. Delivery Rev. 2002, 54, 695-713.
13. L.-T. T. Nguyen, E. J. Vorenkamp, C. J. M. Daumont, G. ten Brinke, A. J. Schouten, Polymer
2010, 51, 1042-1055.
14. V. V. Namboodri, R. S. Varma, Tetrahedron Lett. 2002, 43, 1143-1146.
15. P. Dimitrov. A. Porjazoska, C. P. Novakov, M. Cvetkovska, C. B. Tsvetanov, Polymer 2005,
46, 6820-6828.
16. D. Taton, A. Le Borgne, M. Sepulchre, N. Spassky, Macromol. Chem. Phys. 1994, 195, 139148.
17. S. Mendrek, A. Mendrek, H.-J. Adler, W. Walach, A. Dworak, D. Kuckling, J. Polym. Sci.,
Part A: Polym. Chem. 2008, 46, 2488-2499.
25
18. M. Gauthier, L. Tichagwa, J. S. Downey, S. Gao, Macromolecules 1996, 29, 519-527.
19. S. Rangelov, B. Trzebicka, M. Jamroz-Piegza, A. Dworak, J. Phys. Chem. B 2007, 111, 1112711133.
20. S. Kawaguchi, G. Imai, J. Suzuki, A. Miyahara, T. Kitano, K. Ito, Polymer 1997, 38, 28852891.
26